00:00Precision delivery of medicine
00:02entertainment franchise games absolutely
00:05exploding small modul reactors and the
00:07nuclear Renaissance plus AI moving into
00:11very complex workflows now these were
00:14just a few of the major Tech innovations
00:16that Partners at a16z predicted last
00:19year and our partners are back we just
00:21dropped our list of over 40 plus big
00:24ideas for 2024 a compilation of critical
00:27advancements across all our verticals
00:30from Smart energy grids to Crime
00:31detecting computer vision to
00:33democratizing Miracle drugs like gp1s or
00:36even AI moving from blackbox to clear
00:40boox you can find the full list of 40
00:42plus Builder worthy Pursuits at a16
00:46z.com bigideas 2024 or you can click the
00:50link in our description below but on
00:53Deck today you will hear directly from
00:55one of our partners as we dive even more
00:57deeply into their big idea what what's
01:00the why now what opportunities and what
01:02challenges are on the horizon and how
01:04can you get involved let's dive
01:07in as a reminder the content here is for
01:10informational purposes only should not
01:12be taken as legal business tax or
01:14investment advice or be used to evaluate
01:16any investment or security and is not
01:18directed at any investors or potential
01:20investors in any a6c fund please note
01:24that a6z and its Affiliates may also
01:26maintain investments in the companies
01:27discussed in this podcast for more
01:30details including a link to our
01:31investments please see
01:38disclosures hi I'm Kimberly tan um I'm
01:41an investing partner at Andrus and
01:43Horwitz on the American dynamism and the
01:45Enterprise team and this is my big idea
01:48for 2024 I believe that in 2024 we'll
01:51likely see new applications of computer
01:53vision and video intelligence in the
01:55real world I think leveraging insights
01:57from video data has become pretty Common
01:59Place place in the Enterprise to help
02:01companies make better informed business
02:03decisions but this capability could be
02:05even more powerful in the real world
02:07given how the rich and comprehensive the
02:10nature of the available data is however
02:12many Industries today still lack modern
02:15systems to actually capture and make
02:16sense of that video and oftentimes
02:19customers actually have no existing
02:21video infrastructure or use pretty
02:22Legacy video systems that are difficult
02:25to integrate if you actually want to use
02:26modern software so we see businesses
02:29tackling this problem by leveraging a
02:31hardware and software model allowing
02:33them to sell both the hardware video
02:35cameras as well as the software to
02:37customers these businesses often tailor
02:39the go to market approach to Target a
02:40very specific customer and best serve
02:42those particular needs so for instance
02:45companies like Flock safety examples of
02:47this hardware and software model where
02:49they've built large businesses around
02:51keeping residential neighborhoods and
02:52schools safe um respectively for example
02:55and we think that the same success could
02:57be found in other Industries such such
02:59as for example Transportation
03:01Industrials agriculture mining all right
03:04Kimberly so this in a way is a whole new
03:06Twist on the idea of software is eating
03:08the world in this case it's software is
03:10eating the physical world in a way so
03:13maybe you could just give us a sense of
03:15just how much video is actually
03:17currently being captured yeah definitely
03:19and by the way I think software eating
03:20the physical world is a great
03:22articulation of the broader American
03:23dynamism thesis that we have at the firm
03:25that we're really excited to invest
03:27behind I would say there's there's a
03:30amount of video data that is being
03:32captured just given the general
03:33prevalence of cameras either on
03:35smartphones or just um in a lot of
03:37places that we live our day-to-day lives
03:39and I think the real question is how
03:41much of that data is actually being
03:43captured and utilized in some
03:44interesting way versus just you know
03:47passively existing on different devices
03:49and not actually being analyzed or
03:50processed in any way yeah I think that
03:52makes sense and I think we're all very
03:53familiar with how much data each of us
03:55individually is capturing but what
03:58unlocks in the hardware or the software
04:00to process the data actually allow us to
04:02better leverage this video and um also
04:06maybe have contributed to just how much
04:08data is being collected I would say like
04:09the reason why I'm particularly excited
04:11about um this thesis uh in the next year
04:15or so is I feel like it's the Confluence
04:17of a lot of things that are finally good
04:19enough that were always on the cusp of
04:21being possible but now um have hit some
04:24Cliff that we can see really large
04:26Venture scale outcomes so for example um
04:29like the proliferation of cameras and
04:31availability of videos there's just more
04:33cameras getting installed for different
04:34use cases where the insights aren't
04:36being leveraged as widely they could so
04:38like I mentioned either in people's
04:39smartphones there's all these home Ness
04:42cameras that are getting built out um
04:44there's cameras in people's workplaces
04:46traffic cameras Etc um and I think the
04:49availability of actually just even
04:51having the data there is one thing that
04:53we're reaching a Tipping Point on the
04:55second is I would say the tech just
04:57becoming broadly more uh cost efficient
05:01and uh available so first through like
05:04Cloud cost dropping such that you can
05:06actually do more computation in the
05:07cloud which has historically been a
05:09barrier given just how much data there
05:12is residing in video and how expensive
05:14it is to process that data um and then
05:17secondly on that point like increases in
05:19Edge Computing capacity through cheaper
05:21chips and camera quality per dollar
05:24allowing you to run more complex and
05:26larger models on better cameras at low
05:29cost and thus reducing any sort of band
05:32with constraints you would have had in
05:34actually sending it to the cloud so I
05:36think the tech becoming more efficient
05:38is um a second reason why we're really
05:40excited uh the third would be like just
05:43broadly Innovation on the models
05:44themselves so everybody I feel like now
05:47knows Transformers and how exciting it's
05:49been for like generative AI innovations
05:51that we're seeing in things like chat PT
05:54um and while Transformers were designed
05:55for language tasks um initially they've
05:58actually been applied in Vision um
06:01applications as well and works quite
06:03well for things like image
06:04classification and object detection um
06:06versus the status quo previously so I
06:09think seeing Transformers applied in the
06:11vision space will be really exciting um
06:13just see through using something like
06:15generative AI where you can qu natural
06:17language instead of having to um write a
06:19bunch of code to actually look through
06:20it and understand it in a meaningful way
06:22totally I I also think there's just like
06:23a business model understanding question
06:25which a lot of people thought for a long
06:26time that selling into these difficult
06:28verticals was tough um and I think now
06:30that we have companies like Flock safety
06:32proving that it is possible um people
06:35can apply those lessons into other
06:37vertical where it hasn't currently been
06:38applied yet absolutely so it sounds like
06:41there is an element of the cost curve
06:43changing elements of innovation and
06:45specifically llms changing the game and
06:47maybe through that business models
06:49evolving could you speak to maybe some
06:51of the applications that you think could
06:53emerge here I mean you mentioned
06:54transportation agriculture Industrials
06:57and Mining give us a sense of the types
06:59of applications that might Sprout from
07:02these different factors coming
07:04together yeah um I want to preface by
07:07saying there's probably like way more
07:08than we can even Envision so these are
07:09just examples that I think are exciting
07:12but in transportation for example
07:13there's a lot of back and forth that
07:15happens with just essentially confirming
07:18that you're giving the right item to the
07:20right person at the right time and
07:21charge the correct amount and today a
07:24lot of that is done through very manual
07:27emails invoice processing reconcil a
07:29confirmation of receipt of items
07:31compliance on specific tariffs Etc and a
07:35lot of that through computer vision you
07:37can identify that this was the item that
07:38we gave to this person at this time and
07:40just cut out a lot of that
07:42inefficiency um I think Industrials and
07:45um like other um like pretty large scale
07:49Industries like that there's a lot of
07:51Labor or OSHA compliance on workplace
07:54safety and conditions that have to be
07:55met that inspectors come by you often
07:58have to fill out all these checklists
08:00and make sure and through computer
08:01vision none of which requires actually
08:03monitoring the people themselves you can
08:04make sure that the conditions are safe
08:06for the folks who work there um and then
08:08in something like agriculture you know
08:10the bread and butter of that industry is
08:12livestock or crops or whatever you're
08:13growing and I think there's a lot of
08:15insights that you can be gained without
08:17having to have somebody manually looking
08:19at it all the time about making sure
08:20that the weather conditions are are
08:23correct or the crops are healthy and
08:25such yeah I think that makes a lot of
08:27sense and I think people can easily
08:28imagine those use cases but what about
08:31the use cases where this video is deeply
08:34ingrained in our everyday lives I think
08:35the natural question as it relates to
08:38that scenario is how we avoid the kind
08:42of intelligence that is a slippery slope
08:45into you know a big brother scenario or
08:48something like that how do you see or
08:50how would you calm those concerns or how
08:52do you see regulation potentially
08:53playing a role in this Maybe video
08:56Centric world that you're describing
08:59yeah I think this is a really important
09:00question and topic that all companies in
09:02the space and we monitor pretty closely
09:05um I think privacy should and will
09:08continue to be a very important question
09:10and I think like it it's hard to say how
09:12regulations will shake out but there's a
09:14lot of things that companies can do
09:16around both data rights and data privacy
09:19concerns to make sure who actually owns
09:21the data rest there's lots of things
09:23around how long you're actually allowed
09:25to store the video for and even in
09:29computer vision itself like what are you
09:30tracking um are you tracking for example
09:32just the license plate are you tracking
09:34just the position of a certain piece of
09:36equipment at a certain time instead of
09:38tracking any individual person um so I
09:41think there's a lot that can be done to
09:42make sure that people's privacy rights
09:45are protected while allowing us to use
09:47the technology in a way that makes
09:49people's lives better but it is one
09:51where I think companies have to work
09:52very closely with the stakeholders on
09:54both the regulatory as well as the
09:57personal and business side to make sure
09:59that everyone feels comfortable
10:00absolutely and of course this is a big
10:042024 so we spoke a little bit to this
10:06and the different Trends coming together
10:08but why do you think this is the year
10:11where we're really going to see a
10:12Tipping Point in terms of widespread
10:13adoption and maybe you could just speak
10:15a little bit more to what you hope to
10:16see builders um come and work on this
10:19coming year we've talked a little bit
10:21about this before but I think it's just
10:22a lot of things coming together to make
10:25this the right time to do it um both on
10:27the cost side I think there a little bit
10:29of like a labor element of Labor
10:31shortages in a lot of the industries
10:32that we were talking about that maybe
10:34open the door to software Innovation
10:37where people um were not as used to
10:39buying software in the past um I think
10:42that the general focus that people have
10:45on AI today naturally lends itself to
10:47the question of where else can AI be
10:49apply that isn't um what we've been
10:52seeing with all the exciting generative
10:54creative stuff um and thinking about how
10:56these Technologies can actually work in
10:58these more like brick-and mortar
11:00physical use cases um and then I think
11:02yeah like a lot of companies have now
11:04grown up to be successful companies that
11:06have employed this model and a lot of
11:08new Founders who are starting companies
11:10looking to see which companies they
11:12admire will have these companies as
11:14models to um potentially emulate or take
11:17lessons from whereas that wasn't true a
11:19couple years ago all right I hope you
11:21enjoyed this big idea we do have a lot
11:23more on the way including a new age of
11:26Maritime exploration that takes
11:28advantage of a and computer Vision Plus
11:31AI first games that never end and
11:33whether voice first apps May finally be
11:35having their Moment by the way if you
11:37want to see our full list of 40 plus Big
11:40Ideas today you can head on over to
